Ah, I see, sorry for missing your point. In the paper they injected particles of size "9.55 µm, 1.14 µm, 0.293 µm", and it looks like a red blood cell is 8um. So a purely mechanical filter that got red blood cells would retain the largest of those particles but potentially not the rest. But plasmapheresis uses a combination of centrifugal separation and filters so may be more or less effective than that depending on the properties of the cells and plastics.